ADDITIONAL SYSTEM INFORMATION :
Only affects Windows

A DESCRIPTION OF THE PROBLEM :
Hello,

Following the discussion in https://github.com/openjdk/jdk21u-dev/pull/510 and https://github.com/openjdk/jdk/pull/14523#issuecomment-2096500612, I'm opening a bug report for this stack overflow crash I'm getting in the windows crash reporting.

Tl;dr, my code is causing JVMTI to break, but when the crash is caught by the JVM, a stack overflow happens in the error reporting system, hiding it. A PR fixed a stack overflow there in Java 22+ (and has been backported recently to Java 21), but I'm still getting it on Java 22 and even 23.

So there could be 2 issues here:
- JVMTI crashing on something that was working in past versions
- That crash causing a stack overflow later on

Thanks!

Hope that helps 🙏

REGRESSION : Last worked in version 17.0.11

STEPS TO FOLLOW TO REPRODUCE THE PROBLEM :
Build

EXPECTED VERSUS ACTUAL BEHAVIOR :
EXPECTED -
Program should run properly and not crash
ACTUAL -
Program crashes with a stack overflow

---------- BEGIN SOURCE ----------

---------- END SOURCE ----------

CUSTOMER SUBMITTED WORKAROUND :
A workaround is included in the repo, by not calling FindClass() while internal JDK classes were loaded the crash doesn't happen anymore

FREQUENCY : always